I believe that is accurate and complete. Note that the ranges you named 00000 must be numeric and have no lower or upper limits.dserrano5 wrote:Is the above accurate/complete? Am I forgetting/neglecting/misunderstanding something?
We have the same problem in Portugal. In fact, there are a few places where only the complete postal code (0000-000 format) along with the "city" is enough to know exactly where it is. There are even some cases where almost the complete address would be necessary. In these few and negligible cases you'd need <Hamlet 0000-000 POSTAL-TOWN> to spot the place.dserrano5 wrote:Now, if I look for overlapping or conflicting entries I easily find this example in the codes_germany.txt file:
Code: Select all
Group=Berliner Bezirke (Districts of Berlin)
// PLZ 10119,10179,10247,10367 [...] are all within more than one district
10119=Pankow (Prenzlauer Berg/Weißensee/Pankow)
What should happen to a note entered in 10119? Is it to be assigned to both districts? (which affects the totals and percentages) To none of them? (ditto) To a random one? (yielding different results depending on the phase of the moon).
I'm sure I'll find more interesting cases…
NIG has never allowed the 7-digit postal code, most likely because of the hyphen. It has also never allowed a format like
00000,00000[all other cases]=District2
So what we did, both on EBT city-management assignment rules and on our NIG file codes_portugal_distconc.txt was to assign as much "cities" as possible when in doubt. This yields enormous lists. Here's a very short example:
2100;Santana do Mato=Santarém
It would be lovely, and less prone to errors, to be able to write this as (something like):
That wouldn't be enough, though. Without that and the complete 7-digit postal code, including the hyphen, we will always have the "10119" problem, though. I have no idea if the Germans can solve their problem the same way (they might if, for example, their postal code system is like the US one, getting down to 00000-0000).